A work of pure genius! - Brilliant "Revetec" Engine

[Matra et Alpine]

You wanted to be high and mighty

No I only pointed out that the sensor is a complex part in a complex control system.

and point out that the output of narrow band lambda sensor is a voltage

The output of wideband is a voltage too

... and tried to imply that they don't measure oxygen concentration directly but through the use of a complex multi-variable model. I simply pointed out how a broadband lambda sensor works (since they would probably be using a broadband instead of a narrowband) and that there is a well known direct relation between the voltage or current output and the gas oxygen content.

You yourself listed the parts and all I was saying earlier is that a variable in the engine operation that does not fit in to the model design causes unknown results.
Widebands are sensitive to exhaust gas pressure -- you should know that if you know about them Also the control loop on the pump timing can make a difference in different circumstances.

[hightower99]

No I only pointed out that the sensor is a complex part in a complex control system.

its amazing that when you want to be humble you can. Anyways the sensors are actually quite simple (although they are prescision sensors) and they don't neccessarily need to be part of a complex control system. That is all that I was trying to say.

The output of wideband is a voltage too

Nope broadband lambda sensors output is a current (normally mA), as the voltage is maintained in both the pump and the concentration sensor at a stable value.

You yourself listed the parts and all I was saying earlier is that a variable in the engine operation that does not fit in to the model design causes unknown results.

Again amazing how humble you can be when you want to. While it is true that if they used a lambda sensor that for some reason is not suitable for the current task it will cause unknown effects, hopefully they can find the right lambda sensor for the job. However I would like to point out that we discussing how Lambda sensors are fooled neither of us has implied that the sensor is in perfect working order and that the measured results are to be trusted.

Widebands are sensitive to exhaust gas pressure -- you should know that if you know about them Also the control loop on the pump timing can make a difference in different circumstances.

Yes but the majority of broadband lambda sensors can operate well through a large range of pressures (from idle on a open exhaust engine to full throttle on a turbocharged engine). You do have a point about the control system for the pump timing but as it is trying to keep the sensor cell at lambda 1.00 they basically have a built in diagnostic system (if the sensor cell is not maintained at Lambda 1.00 then there is a problem).

[Matra et Alpine]

Nope broadband lambda sensors output is a current (normally mA), as the voltage is maintained in both the pump and the concentration sensor at a stable value.

Another domenstration of book-knowledge ?

As used in race engines and prgoammable race ECUs
Note the OUTPUT.
You aren't considering the whole component and it's embeeded control loop

[revetec]

We are using wideband for programming and then narrow band for control.

Hightower: I know how a Lambda sensor works and we didn't expect the sensor to read in the way it does. But purely it does give a reading as I have described. This trait has been consistent on multiple lambda sensors. Motec set up our first engine management system in 2001 and it was then we all discovered that the Lambda was showing the combustion effectiveness rather than the actual air/fuel ratio. We spent a lot of time trying to work out why we were getting such a reading, because we actually want to know the actual air/fuel ratio of what we are using, but as it turns out we can only use it as a tuning guide.

In reference to piston height. At 4,000rpm let's say at an ignition point of 30deg BTDC a piston is 10mm from the top of the stroke, our piston is about 5mm from top. This means that the firing volume is half it would be normally.

We have found that our combustion fires richer than normal with a normal injection amount. If we use a normal type injection map the engine chugs black smoke, poorly performs and the Lambda is off the scale in the rich side at higher RPMs. Note: We are injecting roughly the same amount per cycle as we were at 2,000rpm which we had a Lambda reading of 14.7:1. Normally at 4,000rpm we would be injecting slightly more to reach a 14.7:1 ratio (I'm sure you have seen a normal injection map). We had to lean the mixture off to about half of the injection amount to bring the Lambda down to read correctly.

As an example on our RHL4 engine (full throttle) at 2,000rpm our injection was 12ms and at 4,000rpm was 7ms. Fuel pressure was consistent, manifold vacuum was consistent, power output was almost double while torque remained reasonably consistent and at both RPM ranges we had a Lambda reading of 14.7:1. The fuel usage was measured by weight and the engine was using roughly the same amount of fuel in both rev ranges. This was tested in an independent test facility.

We came to only one conclusion from this test. The fuel per firing cycle reduced as the revs increased, although the Lambda was reading the same.

Note: We have changed the piston position profile in the latest tests.

What is your view on this data? And what is anyone's view on this data? Input would be much appreciated.

[Matra et Alpine]

What cylinder wall temperature do you see ?

[revetec]

I'm not sure of the wall temperature as we don't have the equipment to check it. I can say that our exhaust manifold temperature is quite low under full load being about 280degC next to the head. Got to go, I've got to remount the X4v2 back onto the dyno. Cheers

[hightower99]

Matra your graph of a WEGO AFR Output does not say that it is for a broadband Lambda sensor. Are you sure that it isn't a narrow band? Another odd thing is that that graph shows linear output but hopefully you know that Lambda sensor outputs are not even close to linear.

Revetec: Thank you very much for your information. To be honest it boggles my mind if this is found to be true and not some faulty reading from one of the many sensors. I believe I have understood the facts so far.

At 2000RPM and WOT your engine is being fed (n) Lbs/min of fuel by a 12ms injector pulse (per intake stroke per cylinder I assume?). The Lambda sensor is showing Lambda 1.00 (14.7:1 A/F ratio). (n) HP and (n) torque is being produced as measured by a dynamometer.

At 4000RPM and WOT your engine is using an injector timing of 7ms which means it is being fed with (n*1.167)Lbs/min of fuel. The Lambda sensor is still showing Lambda 1.00 even though the engine is being fed at Lambda 1.71(25.1:1 A/F ratio). Roughly (n*2) HP and (n) torque is being produced as measured.

Now the problem is that alot of free gas oxygen (ie O2) is going into the engine while only alittle is coming out.

A question:

Are you running a catalytic converter?

(I am guessing that you are not).

You also say that the exhaust temp is low. (Only 280degC on the exhaust manifold close to the head).

My guess is that you need to hook your engine up to an exhaust analyzer. I would wager that your engine is producing alot of NOx. This would explain the Lambda 1.00 readings (as Lambda sensors don't count/can't sense oxygen bound in NOx). This would also help explain the low exhaust temps (NOx production is an endothermic process).

Have you considered this?

[Matra et Alpine]

Matra your graph of a WEGO AFR Output does not say that it is for a broadband Lambda sensor. Are you sure that it isn't a narrow band? Another odd thing is that that graph shows linear output but hopefully you know that Lambda sensor outputs are not even close to linear.

Thank you, you've finally grasped the POINT
It IS wideband and the output IS linear ... why ? Because BUILT IN to the sensor is an op-amp and feedback. It controls the temp, the pump voltage and normalizes the ouptut. It's as I suspected, you don't see the complexity and the "model" in the o2 sensor and hence dont' grasp why it may be reading figures aligned with "normal" engine operation.

on the left is the "raw" sensor and on the right is the control loop producing the usable output. So a closed loop feedback control system. Which can often have a null or chaotic point where output isn't modelled correctly from the input. From Brad's info, I'd wonder if the sensor heater is up to the job of heating it up all the time as usually the sensor heater is only needed at initial start and for minor heat changes, not major continuous heating.



Good point about the NOx though. Brad ? you got numbers for exhaust contents ??

[revetec]

Revetec: Thank you very much for your information. To be honest it boggles my mind if this is found to be true and not some faulty reading from one of the many sensors. I believe I have understood the facts so far.

At 2000RPM and WOT your engine is being fed (n) Lbs/min of fuel by a 12ms injector pulse (per intake stroke per cylinder I assume?). The Lambda sensor is showing Lambda 1.00 (14.7:1 A/F ratio). (n) HP and (n) torque is being produced as measured by a dynamometer.

At 4000RPM and WOT your engine is using an injector timing of 7ms which means it is being fed with (n*1.167)Lbs/min of fuel. The Lambda sensor is still showing Lambda 1.00 even though the engine is being fed at Lambda 1.71(25.1:1 A/F ratio). Roughly (n*2) HP and (n) torque is being produced as measured.

Now the problem is that alot of free gas oxygen (ie O2) is going into the engine while only alittle is coming out.

A question:

Are you running a catalytic converter?

(I am guessing that you are not).

You also say that the exhaust temp is low. (Only 280degC on the exhaust manifold close to the head).

My guess is that you need to hook your engine up to an exhaust analyzer. I would wager that your engine is producing alot of NOx. This would explain the Lambda 1.00 readings (as Lambda sensors don't count/can't sense oxygen bound in NOx). This would also help explain the low exhaust temps (NOx production is an endothermic process).

Have you considered this?

You are right in your explanation. It is strange that the Lambda sensor is operating this way.

In the report we quoted the wide band high speed sensor from our dyno. We do close loop under 50% load from our narrow band but under the test of course are at 100% load.

We had not used a catalytic converter. We have a 5 gas analyzer measuring HC, CO, CO2, O2 and NOx which are pretty consistent with an engine with no emission control. NOx is usually most prevalent at cruise when mixtures are lean and a high ignition timing is used causing high combustion temperatures, but we experience no benefit in advancing the timing past what we are using. We are not using an EGR valve to reduce NOx.

Our engine likes a higher ignition timing from idle and it varies little during all operational conditions. Usually there is only 10degrees variation throughout all operational conditions. Our test figures on the X4 were produced with a fixed timing of 24deg for the whole dyno ramp. Advancing and retarding from this figure dropped performance slightly (we tested +/-5 deg) The whole engine management programming is totally different from a conventional engine.

From Brad's info, I'd wonder if the sensor heater is up to the job of heating it up all the time as usually the sensor heater is only needed at initial start and for minor heat changes, not major continuous heating.

Our dyno readout only displays when the sensor is at operating temp (displays "cold" when warming up) and the heater is always on when testing is on. This is part of the dyno, and the heater cannot be turned off while displaying.

[hightower99]

In the report we quoted the wide band high speed sensor from our dyno. We do close loop under 50% load from our narrow band but under the test of course are at 100% load.

As I expected

We had not used a catalytic converter. We have a 5 gas analyzer measuring HC, CO, CO2, O2 and NOx which are pretty consistent with an engine with no emission control.

I am guessing you have low HC and CO emissions CO2 should be about normal (for untreated exhaust) NOx should be on the highish side? (maybe not OTT but definately you should still be seeing a good deal at WOT as you are running very lean). Does the O2 reading from the gas analyzer agree with the lambda readings? This is a great way to check the validity of the lambda sensor. If the gas analyzer agrees then gas oxygen is being used to produce something else and the Lambda sensor is working properly. If it doesn't agree then it could mean two things. If it reads more O2 then the lambda sensor measured then the Lambda sensor has a problem, if it reads less O2 then the lambda sensor then there is a really big problem.

Our engine likes a higher ignition timing from idle and it varies little during all operational conditions. Usually there is only 10degrees variation throughout all operational conditions. Our test figures on the X4 were produced with a fixed timing of 24deg for the whole dyno ramp. Advancing and retarding from this figure dropped performance slightly (we tested +/-5 deg) The whole engine management programming is totally different from a conventional engine.

the fact that you engine likes relatively advanced timing is not surprising as it is firing in a more dense mixture. The fact that timing doesn't vary much is overly surprising either as you are only going from idle to just over 4000RPM at WOT. It is odd that you seem to have hit the optimum timing and that you can run the whole rev range well with just that timing.

Our dyno readout only displays when the sensor is at operating temp (displays "cold" when warming up) and the heater is always on when testing is on. This is part of the dyno, and the heater cannot be turned off while displaying.

I doubt that the heater is inadequet but it is something to test because if the sensor isn't hot enough then the readings can be all kinds of wacky.

Also I have been thinking about NOx in your engine. It is produced due to high combustion temperatures but if enough is made it will lead to cooler exhaust temps. However while traveling in the (relatively) cool exhaust, NOx performs an exothermic process where it reverts back to O2 and N2. However this does take some time so it would probably occur close to the end of the exhaust system. I wonder if you put in an exhaust temp probe close to the head and one closer to the end of the exhaust system if you will find that you exhaust starts cooler (than your average piston engine exhaust) when measured close to the head but ends slightly hotter (than your average piston engine exhaust) when measured close to the end of the exhaust system. Note: I am not saying that the actual temperature at the exhaust exit is higher than at the manifold but that your exhaust temperature doesn't drop as much as for a typical engine.


Matra: The raw output from the sensor (ie the input that the output is made from) is the pump current. I agree that there could be a null point but I doubt that that is what is causing the weird readouts.

[jrobson]

ummmm.... bump for the ever elusive BSFC figures....

[revetec]

For your viewing pleasure! Cheers, Brad

EDIT: I've added it as an attachment Brad, most of us were only seeing a red cross MetA

Cool! I forgot how to add images. hehhee

[3yearsharehold]

Revetec
So you have more power than a 1.6 lt vw...mmmmm...and you have a 2.4 lt and !

[revetec]

Revetec
So you have more power than a 1.6 lt vw...mmmmm...and you have a 2.4 lt and !

No...More torque per litre in the lower rev range.

At 25% throttle and 1,500rpm with our 2.4litre (about 50hp), has better acceleration than the high performance VW 1.8litre engine at 100% throttle and 4,000rpm (about 200hp) BTW. We are producing 183Nm@1,500rpm and 187Nm@3,500rpm and peaks at 202Nm@3,000rpm (remember this engine was set up as a test platform for aviation with a normal operating RPM of 3,000rpm), so the acceleration off the mark is hard and reasonably consistent as the revs increase. We also have quite high torque at low throttle response as well.

Did you actually look at GTM Director Shane Subloo's Interview after the test day on our website? Go look at it and watch what he says in his comparison. Gallery Video

No matter what you achieve there will always be a knocker.

Next phase...Certifying our BSFC figure independently (next month) but I suppose, even if it shows it being great as we have performed in-house, there will always be someone there to knock it.

I will be away over the next month while independently certifying power, torque, BMEP and BSFC. I wont be available to comment on here until after certification. Cheers

[hightower99]

At 25% throttle and 1,500rpm with our 2.4litre (about 50hp), has better acceleration than the high performance VW 1.8litre engine at 100% throttle and 4,000rpm (about 200hp)

A few questions:

To make 50HP at 1500RPM the engine needs to be making 175lbs-ft. of torque... Are you claiming that you engine is capable of that at 25% throttle? (meaning it is capable of over 500lbs-ft. of torque if you used WOT?)

Are you claiming that the engine you are comparing the revetec to is a 1.8L VW engine? capable of 200HP+?

In order for the VW engine to make 200HP at 4000RPM it needs to be making 263lbs-ft. pf torque... which means it is making considerably more torque per unit displacement then your revetec engine (146lbs-ft./L for VW vs. 73lbs-ft./L for Revetec).

Please explain how 50HP can achieve "better acceleration" than 200HP?

methinks you made that post without thinking too much...

BTW. We are producing 183Nm@1,500rpm and 187Nm@3,500rpm and peaks at 202Nm@3,000rpm (remember this engine was set up as a test platform for aviation with a normal operating RPM of 3,000rpm), so the acceleration off the mark is hard and reasonably consistent as the revs increase. We also have quite high torque at low throttle response as well.

That's odd...
That means that your engine is making:
39HP @ 1500RPM at WOT
92HP @ 3500RPM at WOT
85HP @ 3000RPM at WOT

So how can it make roughly 50HP at 1500RPM and 25% throttle???
And how can it even come close to the same performance of a 200HP+ engine?

Next phase...Certifying our BSFC figure independently (next month)

I can't wait to see the results

I will be away over the next month while independently certifying power, torque, BMEP and BSFC. I wont be available to comment on here until after certification. Cheers

I hope you also spend time figuring out what exactly was happening with the weird Oxygen sensor readings.... But maybe you forgot about them?